Application portfolio: From a reactive inventory to an AI-driven strategy

Today, the average large enterprise runs 2,191 applications and has no idea what half of them do (Torii, 2026). Forty-eight percent of enterprise applications are unmanaged (Productiv, 2024). Shadow IT accounts for between 30% and 40% of IT spending (Gartner). This is not a governance crisis. It is a strategic opportunity that few CIOs have yet seized.

The application portfolio: the blind spot of IT performance

For years, the priority was delivery. Delivering projects, deploying tools, and responding to business requests. The result is clear: sprawling IT systems, uncontrolled application stacks, and technical debt that quietly consumes between 15 and 20% of the total IT budget. Without anyone ever really deciding that it was a good idea.

SaaS has amplified the phenomenon. Whereas a software purchase once required a request for proposals and IT department approval, today all it takes is a credit card and a work email address. The result: 65% of SaaS applications in enterprises are not approved by the IT department. And AI is accelerating the trend even further; Torii describes the phenomenon as “shadow AI,” a new invisible layer added to the existing chaos.

IBM has put it in numbers: only 36% of IT executives say their investments in cloud, data, and AI are managed as an integrated portfolio, defined by business objectives and a common architecture. The rest are flying by the seat of their pants.

This is where the opportunity lies. Not in short-term cost reduction, though that is real and substantial, but in the ability to transform a burden into a strategic lever.

From complexity as a burden to clarity as an advantage

Application portfolio rationalization is one of the best-documented exercises in enterprise architecture. Inventory of applications, assessment of their business value and technical health, disposition decisions (retain, consolidate, modernize, retire), and roadmap development. On paper, the process is well-known. In reality, it fails regularly.

Why? Because it relies on incomplete data, time-consuming interviews, political trade-offs, and spreadsheets that become obsolete the very day they’re approved. A typical rationalization exercise takes an average of twelve weeks. For a portfolio of 600 applications, this represents a considerable human investment, with a high risk that the conclusions will be contradicted by on-the-ground reality six months later.

This model is currently undergoing a profound rethinking. Not because the method is flawed, but because the available tools have changed.

AI as a Compass for Rationalization

Artificial intelligence is transforming application rationalization in three decisive ways.

First, speed. Cognizant measured this across dozens of client projects in 2024 and 2025: an AI-driven approach cuts the duration of the assessment phase in half: from twelve weeks to six.

The agents automatically aggregate data from management systems, identify functional redundancies, flag inconsistencies, and generate preliminary analyses that teams need only validate and contextualize.

Second, accuracy. Whereas the traditional approach relies on usage declarations, often biased by internal political issues, AI analyzes actual application usage patterns. A tool deemed “critical” by one department may involve only three active users per week. Another, deemed secondary, may turn out to be central to ten cross-functional processes. It is this level of granularity that AI makes possible at scale.

Third, continuity. One of the historical limitations of rationalization is that it produces a snapshot, relevant at a given moment, but outdated as soon as the IT system evolves. AI enables a shift from a one-off exercise to a continuous portfolio monitoring program: real-time detection of cost overruns, new redundancies, orphaned applications, and compliance risks. Rationalization is no longer a project; it becomes a living governance practice.

Cognizant has documented the case of an organization with over 600 applications that reduced its annual IT spending by $13 million using this approach, by identifying redundant assets, unused licenses, and consolidation opportunities that teams previously lacked the visibility to see.

Modernizing Without Rebuilding Everything

Streamlining isn’t just about shutting down applications. It involves defining a path forward for each one. Practitioners typically use a framework with five to seven options: keep as-is, overhaul (application modernization), replace with an off-the-shelf solution, consolidate with another existing application, migrate to the cloud, or retire permanently.

The decision is not purely technical. It is strategic. An older application, costly to maintain, may be at the heart of a differentiating business process that no off-the-shelf solution can replicate. Retiring it would be a mistake. Modernizing it is a priority. Conversely, five project management tools deployed across five different departments represent an obvious opportunity for savings and standardization.

This is where enterprise architecture plays its structuring role: aligning applications with business capabilities, data flows, regulatory requirements, and the strategic trajectory. Without this systemic vision, rationalization is reduced to a hunt for savings: useful, but insufficient.

The Hidden Value of a Streamlined Portfolio

A streamlined application portfolio isn’t just less expensive. It’s more agile. This is the benefit IT teams most often cite after a successful initiative: the ability to deploy new capabilities faster, integrate partners and suppliers more easily, and build AI on a solid foundation.

For the key lesson of 2024–2025 is this: AI is difficult to integrate into a chaotic IT infrastructure. It requires high-quality data, interoperable systems, and clear governance. Streamlining the tech portfolio is therefore not merely an administrative prerequisite for AI, it is the condition for its operational success.

At Gabriel Greenfield, our Meridian approach - Suggest, Facilitate, Support - naturally applies to this process. Suggesting the trajectory for each application asset based on objective analysis. Facilitating trade-offs between the CIO, business units, and senior management, which are often more political than technical. Supporting implementation over the long term, because an application roadmap is only valuable if it is actually executed.

Tech portfolio rationalization is not a cost-cutting exercise. It is an exercise in preparing for the future.

And in a world where AI is rapidly reshaping architectures, it has never been more urgent, or more profitable, to know what you truly own.

REFERENCES

• Torii — 2026 SaaS Benchmark Report: average company = 2,191 applications; 61% not approved by IT

• Productiv — Enterprise Application Report (2024): 48% of enterprise applications are unmanaged

• Gartner — Shadow IT represents 30–40% of large enterprise IT spending

• IBM Institute for Business Value — Only 36% of enterprise tech executives manage cloud/data/AI as integrated portfolios

• Cognizant — Optimizing Application Portfolio Rationalization with AI (October 2025)

• Journal du Net / Programmez — Technical debt accounts for 15-20% of the annual IT budget

• Gartner — SaaS spending: 41% of software spend in 2023, up from 21% in 2019